@@NoKingsNoPeasants lol i mean it was arledy like that,

just started with Python again. When it comes to learning to code you really do need to just code alot and have the persistence to stay with a difficult task for possibly a long time. So I do that. But, I found that if (after to much time) I cant find a specific solution to a problem and have a hard time outlining the solution I might ask chatGPT to: 1. Give me an outline (in plain english/swedish) of the solution to this probelm, without providing any code. Then I proceed with the task at hand. If i still cant figure it out I might ask it to give me a solution using only psuedocode so that I still need to actually code it in python and practice syntax and builtin functions, etc. I've also used it to learn more about Big O notation (your video on this is amazing). chatGPT seem to be pretty good at giving and explaining the time and space complexity of a program. So I use it to verify that my own idea of the complexity is correct and if it's not I'l keep asking GPT to provide more in depth explanations.

yeah, I let him write all the code lol

helps me a lot its my 24hs teacher

Here's my approach to this subject - I want to make apps without writing a single line of code... The most moronic choice one can make right now,, is investing in learning computer science or coding - it''s like trying to become a life-guard who works by saving dolphins from drownng... Do you want to see how you make coding these days? Look at this - yesterday I got a funny idea of creating myelf an "all-mighty power armor" for the avaliable LLMs. that can turn even the most helpless form of basic GPT-3 into a 'digitally omnipotent demigod of data'. This is how I defined the general premise in form of a prompt given to yet-another coding LLM named Cognosys ... Create an universal embedding framework for autonomous LLM agents with user-frriendly GUI (might be accessible through browser) which adds new capabilities to base model and can be customized to fit the user's requirements. This framework should work with models distributed as *.bin files and should handle multiple different LLMs and allow them to be downloaded to a local storage. Such embedding framework should include such modules and servives like: 1. internet access (with http and API protocols), 2. short- and long-term memory modules integrated with a dynamic and accessible database stored locally (preferred option), 3. ability to utilize 2 different core LLMs simultaneously - larger one responsible for logical operations (thinking) and smaller (faster) one responsible for executing 'direct' operations (digital motor functions). It should also allow to connect/disconnect LLMs in *.bin format 'on the fly' 4. capability to process and operate on files in most common formats (images, documents,music, videos, and formats associated with used scripts and protocols 5. allow uploadig/downloading files from pc or to be accessed via internet 6. ability to write and execute scrits (java, python, C++, etc) 7. documen6t editor 8. be able to display images/documents, a movie/sound player 9. tools allowing to create/modify documents and other files in common formats 10. - your own ideas Here's is a GUI which you can use as good example: aiagent.app/ This one is also ok - github.com/oobabooga/text-generation-webui - I'd love to hsve something like both those frameworks connected... #### It took around 20 minutes to complete the task.

i like feeling like i am talking to someone lol

I like to be polite to chatgpt, he works very hard for me

Jesus

​@@lucabernardini3975 There are no imaginary deities involved.

@@outtabubblegum7034 yeah but when they come and they read my logs to their ancestor i want them to like me okay

nothing insane it's 2023

@@nikitapestov1711 Imagine 2030s. Gonna be a wild and awesome time.

Dude same 👊.

It's a she 😂

@@hahnonimus I asked chat gpt and it said "I don't have a gender, as I'm just a computer program. Is there something specific you'd like to ask or discuss?"

Bro I do the variable naming alot.

Dude, let it come out first.

​@@morgengabe1 lol its like the experience request of 10 years in mojo

not storing conversation history is one issue, but the fact you have now voluntarily provided more data for the model to learn from is where I think companies will be concerned with as well.

++++

Thank you! Hope you enjoy it 🙌

Chatgpt is excellent in teaching you how to code. Just ask ChatGPT any questions about anything in programming language you want to learn it will teach you, and it will explain the code. Always ask it to give you an example or sample code and to explain the code to you like to a 10 year old. Then practice and practice creating your own code and have chatgpt explain your code you wrote and correct it for you. You learn and get better if chatgpt helps you in troubleshooting

You better find another job. Don’t waste you education on a dead end like programming. There is gonna be like 10% of coders who will survive by adopting and they are the ones with tons of experiences and connections to stay in and babysit the new AI. Sorry to dump it like this on you but you will thank me in 10 years or less

@@MD-gk4uh Do not listen to this guy, if you knew anything about the industry, you'd know that 'coding'' is around 15% of the job. Even if coding gets automated fully, thats just a very small part of what we do. Computer Science is a great and very interesting career and the only people being replaced are the ones who give up because they're too scared of AI like this guy.

@@MD-gk4uh bruh

@@strawhatnoah6807 Keep telling yourself that. Software designer/engineer here. Don’t have to take my advice. Learn the hard way.

The legal implications of using GPT at work and using it to generate or modify proprietary code depend on a variety of factors, including the specific use case, the jurisdiction in which the work is being performed, and the terms of use of the GPT platform being used. Generally speaking, if you are using GPT to generate or modify code for your company's proprietary software, you should exercise caution and ensure that you have the proper authorization to do so. You should also be aware of any applicable intellectual property laws that may apply, such as copyright law or trade secret law. In terms of using GPT itself, the terms of use for the platform you are using should outline any restrictions or limitations on how the platform can be used. For example, OpenAI, the creator of GPT, has specific terms of use for its GPT platform that prohibit certain uses, such as using GPT to create deepfakes or to engage in spam or fraud. Overall, it's important to consult with legal experts in your jurisdiction and to carefully review the terms of use of any platform you are considering using before incorporating GPT or any other AI technology into your work.

​@@chillydoog You wrote this with ChatGPT 😂.

@@pkom6418 🤫

​@@pkom6418 lol yep

you can use anything u generate, even on a free plan. openai says so

That's what I do. It would make it easier to search the prompts and completions later because ChatGPT always automatically creates a specific title for each thread based on the first prompt of the thread.

It do be dumb af sometimes. I learned to not give it difficult problems

Python is a very common first programming language, so you don’t need to learn anything before learning it.

I can't even rely on Myself dude 😂. Anyway, it doesn't take long for programmers that start using gpt to figure out the flaws

It's not cheating, without ChatGPT you gonna search it on google any way. But with Chat GPT you save a lot of googling time.

2017 is when I started learning to code(I stopped after a year 😂 ) back then I do alot of reddit,stack overflow and just plain old google to get the answers for the problems I want. Now in 2023 chatgpt has vastly reduced that searching problem.

But it made u click 😉

Just say “continue”

@@TechWithTim I've done this several times, sometimes it rewrites the code and it stops again sometime it continues with a random part sometimes it asks me "hi how can i help you".

@@M7ilan Yeah, happens to me a lot, too. I try to be more specific and limit my prompts to single functions, not include HTML / CSS (if relevant), etc. The less code it has to produce, the better. If the output is longer than the message it can produce in a single go, it always ends up being a drag to try and get it to work.

sometimes after typing continue it starts answering an unrelated question I didn't ask. For example I asked it to refactor code then it stopped halfway through, after typing continue it started giving me productivity tips.....

Context Window has a specific size. I think if you are above this there will be some Kind of compression on your Chat history. Does someone know how exactly this works? I testen something like that with a prompt at the begining like give me a random Number and then give me lorum Ipsum thousands of tokens long. The next question was what was the Number and it gives me Not the right Answer.

Are u using gpt4?

Did u watch the video? I mention exactly that at the beginning

@spacejamzyt6461 y'all are making me scared with all these ai will take our jobs. My future is bleak then

Shhhhhhhh

Milking, maybe, can't blame them. But it's for sure revolutionary

@@StevenMartinGuitar No it's not, its like Simisimi? So basically the only difference is this Ai is train to larger data. And every programmer know that even Tim

@@ImDGreat Never heard of Simisimi before this, but you're right. Compared to it ChatGPT isn't revolutionary at all. Bot: Hello Me: What are you? Bot: your mom Me: You're not my mom Bot: Your mom _____ Me: :( Bot: smile back at me

Lol alright, internet dude. If you don't see how it's a game changer, you're probably not qualified to have an opinion. Do you even understand what "more data" even means? You understand that neural architecture these models are built on didn't exist 6 years ago? It's pretty revolutionary there, champ.

He's another shill for ChatGPT or whatever fart he's sniffing that he thinks will get him likes and subscribes.